I am running a Multinomial Logit model using the glogit option in PROC LOGISTIC. However, I am having some trouble finding a way to print the individual probabilities for EACH group in my DV (there are 3, with 2 as the ref). I know how to run this option in binary, but can't get it to work correctly in multinomial.
What I want is a new data set with three new variables, those being the probabilities of group membership based on the model. Is this possible?
Stated differently, when I run a multinomial logistic model (3 groups, 2 as the reference) and request the output and p='var', what do the probabilities in p represent in this case? The odds of being in group 2?